计算流体力学中的重构方法

Reconstruction is one of the most fundamental and indispensable steps in the development of computational fluid dynamics(CFD). In essence, reconstruction techniques aim to recover or approximate the underlying "true" solution profile from limited discrete information, such as cell-averaged values (volume-integrated averages, VIA) or pointwise values (PV). Over the history of CFD, many numerical methods have been devised to achieve this goal, including the MUSCL scheme, various weighted esscntially non-oscillatory (WENO) schemes, and the THINC scheme, among others.

In addition to approaches that rely solely on VIA or PV, more advanced for mulations have been developed that simultaneously incorporate both types of information, as well as higher-order derivatives. These so-called multi-moment schemes provide enhanced accuracy and flexibility, enabling the construction of higher-order discretizations that are well-suited for resolving complex flow structures.

This talk will highlight some progress in reconstruction techniques, with particular emphasis on how they can be designed to achieve both numerical consistency and generality. Furthermore, attention will be also given to how these methods can be adapted to modern high-performance computing architectures, ensuring that they remain efficient and scalable for large-scale CFD applications.

      肖锋博士现任东京科学大学（原东京工业大学）机械工程系教授。近年来致力于计算流体力学（CFD）、地球物理流体动力学建模、计算物理以及CFD工程应用等领域的研究。已在学术期刊发表论文180余篇，在国际及日本国内会议与研讨会上完成学术报告300余场次。现任日本机械学会（JSME）及日本流体力学学会（JSFM）会士，曾获日本计算力学协会（JACM）2013年度计算力学奖、日本机械学会2011年度计算力学成就奖、以及JACM 2009年会士称号。目前担任国际权威期刊Journal of Computational Physics（JCP）副主编。